Patients with rapid cycling bipolar disorder (RCBD) experience at least four episodes of affective illness (depression, hypomania, and/or mania) in a year. They suffer significant morbidity and are frequently resistant to conventional treatments. Results from Z01 MH 02614-01 CPB show that the mood cycles of patients with RCBD may be associated with shifts in the phase (timing) of their circadian rhythms. If so, then the use of light, dark, or melatonin to shift circadian rhythms (or to increase the amplitude of the nocturnal melatonin rhythm, and thereby make the circadian system more resistant to phase shifts) could have therapeutic effects. In this project, we test the clinical effects of midday bright light (vs. low-dose negative ion generator), melatonin (vs. placebo), and dim (vs. bright) light in patients with RCBD. The studies of midday bright light and of melatonin are both outpatient treatment trials in which the experimental treatment is added as an adjunct to a stable medication regimen. The patient uses the experimental and the control treatment for three months each (in randomized order), with a one-month wash-out period in between. Our work with bright light is the most advanced. Bright light therapy may be a useful therapeutic modality in this population because of its rapid onset and offset of action. Patients can be instructed to use the treatment when they are euthymic or depressed but to forego it when they are hypomanic, thereby allowing them to benefit from light's anti- depressant effects without experiencing increased cycling. In a pilot trial of light therapy in patients with RCBD, we obtained preliminary evidence that morning light therapy exacerbated their cycling, evening light therapy had no effect, and midday light therapy had a mood- stabilizing effect. We hypothesize that midday light therapy was therapeutic because it increased the amplitude of nocturnal melatonin secretion, thereby rendering the rhythm more resistant to the phase shifts that appear to be associated with mood cycling. We are now testing this hypothesis, using low-dose negative ion generator as a control condition. In addition to testing the clinical effects of midday light, we will measure the patient's nocturnal melatonin secretion in each experimental condition to test the hypothesis that midday light therapy increases the amplitude of nocturnal melatonin secretion. The second study in this project consists of a trial of 10 mg of oral melatonin, testing its ability to stabilize mood and to consolidate the fragmented sleep patterns frequently seen in patients with RCBD. Finally, the third experiment tests the acute effects of eight hours of bright vs. dim light (adminstered in randomized order) on hypomania. This study will demonstrate whether light has clinical effects in hypomania, and should provide preliminary evidence as to whether exposure to dim light could have therapeutic effects in hypomanic patients.